Approximately 20% of breast cancers detected through mammography are pre-invasive Ductal Carcinoma in situ (DCIS). If left untreated, approximately 20-50% of DCIS will progress to more deadly Invasive Ductal Carcinoma (IDC). No prognostic biomarkers can reliably predict the risk of progression from DCIS to IDC. Similar genomic profiles of matched pre-invasive DCIS and IDC suggests that the progression is not driven by genetic aberrations in DCIS cells, but microenvironmental factors, such as hypoxia and metabolic stress prevalent in DCIS, may drive the transition. We need innovative models to investigate how to halt steps of DCIS progression to invasive phenotypes and subsequent metastasis from the primary site. This proposal directly addresses this unmet need by developing a novel three-dimensional in vitro organoid model that recapitulates key hallmarks of DCIS to IDC progression: tumor-size induced hypoxia and metabolic stress, tumor heterogeneity and spontaneous emergence of migratory phenotype in the same parent cells without any additional stimulus. A tangible advantage of the proposed organoid models is the ability to precisely and reproducibly study how the hypoxic microenvironment induces tumor migration in real time and in isolation from non-tumor cells present in vivo, providing unique opportunity to define tumor-intrinsic mechanisms of DCIS to IDC progression. Our preliminary observations lead to central hypothesis that tumor size-induced hypoxia establishes a ?hypoxic secretome?, which initiates the migratory phenotype; the hypoxic secretome then cooperate with intracellular signaling networks to independently maintain cell migration. We propose three independent but inter-related aims to link hypoxic secretome with the initiation, maintenance and spatial distribution of migratory phenotypes. Aim 1 will engineer size-controlled DCIS organoids (150-600 m) with controlled hypoxic microenvironments to identify and examine how hypoxic secretome initiates migratory phenotype. We will combine experimental organoid models with time-lapse imaging and computational approaches to study organoid migration. Aim 2 will demonstrate that migratory cells can re-establish the secretome and maintain migratory phenotype independent of hypoxia. We will reconstruct an intracellular signaling network activated by the hypoxic secretome using microarray data. We will verify these gene expression signatures in sorted migratory and non-migratory cells, and validate them using secretome inhibition studies. Aim 3 will investigate, for the first time, the spatial distribution and origin of the migratory phenotype. We will use CRISPR-based gene knock-in (FP-labeling), automated image analyses, and a deep-learning algorithm to track and visualize the emergence of migratory phenotypes from the hypoxic core outward to the periphery or from the migratory front. The successful development of this 3D organoid model and completion of the proposed work will provide answers to two fundamental questions in the progression of invasive breast cancer: 1) What causes some DCIS cells to become migratory and develop into invasive tumors? 2) How and where does the migratory phenotype (IDC) emerge? The mechanistic understanding gained from these studies will improve diagnosis, lead to the development of treatment strategies to arrest invasion at the pre-malignant stage, and thus prevent patient overtreatment. It is straightforward to generalize our system to other tumor types, development of tumor/stromal co-culture, and drug screening.